This repository contains the R scripts to curate, ensure quality control, normalize, and generate final model results from the raw data collected via the S-Interface and other tools in the context of a naturalistic study on deadlines. The study was funded by NSF (grant #1704682) and appeared in the paper Sympathetic Activation in Deadlines of Deskbound Research - A Study in the Wild, published in CHI 2023.
- R and RStudio
- Required packages
Packages are available on CRAN and can be installed using a simple call to install.packages():
install.packages('PackageName')
The Data Curation (dc) scrips will process the raw data and create the model data. The Modeling Scripts (ms) will perform the multy linear regression and produce the visualizations.
- 0.dc-all-script.R
- The only script to run to curate, control the quality, transform and get the final data for modeling. Download raw_and_noise_removed_df.csv from OSF Sympathetic Activation in Deadlines put it on the "all-subj-data" folder under "curated-data" folder.
- This script calls all the following scripts sequentially:
- 2.dc-process-activity-app-usage-data.R
- 3.dc-quality-control-phase-one.R
- 4.dc-generate-transformed-data.R
- 5.dc-generate-normalized-data.R
- 6.dc-merge-all-data.R
- 7.dc-generate-meta-data-break-activity.R
- 8.dc-generate-model-data.R
- ms-descriptive.rmd
- ms-model-visualization.rmd
Details of Utility Scripts (us)
The Utility Scripts (us) are used to define the common functions like read-write files, and are being called from the Data Curation (dc) scripts.
- us-common-functions.R
- Useful functions that are called from almost all scripts.
- us-filter-pp.R
- Removes noise from PP signals. It is called from 1.dc-curate-and-process-data.
- us-down-sample-pp.R
- Downsamples data to 1 fps. It is called from 1.dc-curate-and-process-data.
Details of Data Curation (dc) & Modeling Scripts (ms)
- 1.dc-curate-and-process-data.R
- For each participant the script does the following:
- Reads the original perinasal perspiration signal data, removes noise, downsamples to 1 frame per second (fps)
- Reads the E4 and iWatch signal files, downsamples them, removes noise, and merges them with the PP signal
- 2.dc-process-activity-app-usage-data.R
- Processes and finalizes the participant's activity data
- Processes and finalizes the app usage data
- 3.dc-quality-control-phase-one.R
- Performs filtering on physiological data, to remove the signals with invalid range
- 4.dc-generate-transformed-data.R
- Performs log transformation on all modalities of the physiological channel
- 5.dc-generate-normalized-data.R
- Generates normalized data in respect of the Resting Baseline session
- 6.dc-merge-all-data.R
- Gathers and merges the physiological, psychometrics, activity data for all participants
- 7.dc-generate-meta-data-break-activity.R
- Generates data for total breaks, time of the break, frequency of the breaks etc.
- 8.dc-generate-model-data.R
- Generates the working session mean data to perform the linear modeling
- ms-descriptive.rmd
- Produces the exploratory data visualization
- ms-model-visualization.rmd
- Checks collinearity and performs linear modeling and Visualizes the significant predictors of the model